Windows & Linux GUI for marking bounded boxes of objects in images for training Yolo v3 and v2
To compile on Windows open yolo_mark.sln
in MSVS2013/2015, compile it x64 & Release and run the file: x64/Release/yolo_mark.cmd
. Change paths in yolo_mark.sln
to the OpenCV 2.x/3.x installed on your computer:
(right click on project) -> properties -> C/C++ -> General -> Additional Include Directories: C:\opencv_3.0\opencv\build\include;
(right click on project) -> properties -> Linker -> General -> Additional Library Directories: C:\opencv_3.0\opencv\build\x64\vc14\lib;
To compile on Linux type in console 3 commands:
cmake .
make
./linux_mark.sh
Supported both: OpenCV 2.x and OpenCV 3.x
x64/Release/yolo_mark.cmd
./linux_mark.sh
x64/Release/data/img
.jpg
-images to this directory x64/Release/data/img
x64/Release/data/obj.data
: https://github.com/AlexeyAB/Yolo_mark/blob/master/x64/Release/data/obj.data#L1x64/Release/data/obj.names
: https://github.com/AlexeyAB/Yolo_mark/blob/master/x64/Release/data/obj.namesx64\Release\yolo_mark.cmd
x64/Release/yolo-obj.cfg
:filter
-value
(classes + 5)*5
: https://github.com/AlexeyAB/Yolo_mark/blob/master/x64/Release/yolo-obj.cfg#L224(classes + 5)*3
3.1 Download pre-trained weights for the convolutional layers (76 MB): http://pjreddie.com/media/files/darknet19_448.conv.23
3.2 Put files: yolo-obj.cfg
, data/train.txt
, data/obj.names
, data/obj.data
, darknet19_448.conv.23
and directory data/img
near with executable darknet
-file, and start training: darknet detector train data/obj.data yolo-obj.cfg darknet19_448.conv.23
For a detailed description, see: https://github.com/AlexeyAB/darknet#how-to-train-to-detect-your-custom-objects
To get frames from videofile (save each N frame, in example N=10), you can use this command:
yolo_mark.exe data/img cap_video test.mp4 10
./yolo_mark x64/Release/data/img cap_video test.mp4 10
Directory data/img
should be created before this. Also on Windows, the file opencv_ffmpeg340_64.dll
from opencv\build\bin
should be placed near with yolo_mark.exe
.
As a result, many frames will be collected in the directory data/img
. Then you can label them manually using such command:
yolo_mark.exe data/img data/train.txt data/obj.names
./yolo_mark x64/Release/data/img x64/Release/data/train.txt x64/Release/data/obj.names
/x64/Release/
yolo_mark.cmd
- example hot to use yolo mark: yolo_mark.exe data/img data/train.txt data/obj.names
train_obj.cmd
- example how to train yolo for your custom objects (put this file near with darknet.exe): darknet.exe detector train data/obj.data yolo-obj.cfg darknet19_448.conv.23
yolo-obj.cfg
- example of yoloV3-neural-network for 2 object/x64/Release/data/
obj.names
- example of list with object namesobj.data
- example with configuration for training Yolo v3train.txt
- example with list of image filenames for training Yolo v3/x64/Release/data/img/air4.txt
- example with coordinates of objects on image air4.jpg
with aircrafts (class=0)
Button | Description |
---|---|
Left | Draw box |
Right | Move box |
Shortcut | Description |
---|---|
→ | Next image |
← | Previous image |
r | Delete selected box (mouse hovered) |
c | Clear all marks on the current image |
p | Copy previous mark |
o | Track objects |
ESC | Close application |
n | One object per image |
0-9 | Object id |
m | Show coords |
w | Line width |
k | Hide object name |
h | Help |
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。